Behavior Predicts Genetic Structure in a Wild Primate Group JEANNE ALTMANN*Ts, SUSAN C
Total Page:16
File Type:pdf, Size:1020Kb
Proc. Natl. Acad. Sci. USA Vol. 93, pp. 5797-5801, June 1996 Population Biology Behavior predicts genetic structure in a wild primate group JEANNE ALTMANN*tS, SUSAN C. ALBERTSt§, SUSAN A. HAINESI, JEAN DUBACHt, PHILIP MURUTHItII, TREVOR COOTE¶, ELI GEFFEN**, DAVID J. CHEESMAN1, RAPHAEL S. MUTUTUAt, SERAH N. SAIYALELS, ROBERT K. WAYNEtt, ROBERT C. LACYt, AND MICHAEL W. BRUFORD¶ *Department of Ecology and Evolution, University of Chicago, 940 East 57th Street, Chicago, IL 60637; §MCZ Laboratories, Harvard University, 26 Oxford Street, Cambridge, MA 02143; 1Conservation Genetics Group, Institute of Zoology, London NW1 4RY, United Kingdom; tDepartment of Conservation Biology, Chicago Zoological Society, Brookfield, IL 60513; 1Department of Ecology and Evolutionary Biology, Princeton University, Princeton, NJ 08544; **The Institute for Nature Conservation Research, Faculty of Life Sciences, Tel Aviv University, Ramat Aviv 69978, Israel; tInstitute of Primate Research, National Museums of Kenya, P.O. Box 24481, Nairobi, Kenya; and ttDepartment of Biology, University of California, Los Angeles, CA 90024 Communicated by Sarah Blaffer Hrdy, University of California, Davis, Ca, February 12, 1996 (received for review October 31, 1995) ABSTRACT The predictability of genetic structure from ally peaking in early adulthood (age, 8-10 years) and then social structure and differential mating success was tested in declining more or less rapidly (12, 14, 15). Potentially fertile wild baboons. Baboon populations are subdivided into cohe- matings occur primarily within close sexual consortships, sive social groups that include multiple adults of both sexes. which involve mate-guarding during days of likely conception. As in many mammals, males are the dispersing sex. Social A male can monopolize only a single female when several are structure and behavior successfully predicted molecular ge- simultaneously in estrus. netic measures of relatedness and variance in reproductive We first focused on intensive paternity analysis for a cohort success. In the first quantitative test of the priority-of-access of 27 surviving offspring conceived in Lodge Group (7, 16) model among wild primates, the reproductive priority of from 1985 through 1988, a period for which both behavioral dominant males was confirmed by molecular genetic analysis. and genetic data were available. The time window was then However, the resultant high short-term variance in reproduc- extended both backward 4 years, using only genetic data for 14 tive success did not translate into equally high long-term surviving offspring, and forward 3 years, using only behavioral variance because male dominance status was unstable. An data for 42 conceptions (22 surviving offspring), to permit important consequence of high but unstable short-term vari- estimation of differential reproductive success for 11 years, ance is that age cohorts will tend to be paternal sibships and approximately the duration of adulthood in this species. social groups will be genetically substructured by age. In this study, we combined molecular genetic data with METHODS long-term behavioral and demographic data to examine sev- Behavioral and Demographic Data. Behavioral studies on eral aspects of behavior-genetic relationships that are central Lodge Group began in 1984 (7) and became regular in the next to the evolution of primate social systems. The first of these is year. Behavioral and reproductive data were collected on all the priority-of-access model, which predicts that dominance observation days. Reproductive data include each female's status among adult males determines access to estrous females reproductive condition (including size and turgescence of sex (1) and that variability in the number of offspring fathered by skin, which in baboons can be used to determine days of likely males will, therefore, directly reflect both the males' domi- ovulation to within a few days) and duration of sexual con- nance status and the number of simultaneously estrus females sortships. Agonistic data include records of wins/losses based (2). Second, we investigated the widespread assumption that on aggressive and submissive behaviors, which are used to short-term differences in mating success or paternity success determine male dominance ranks (17). Because the habitat are stable and, therefore, predictive of lifetime differences in was open and group size not too large ('50), both baboons and reproductive success (for review, see refs. 3 and 4). Third, we observers could readily monitor reproductive and agonistic examined the hypothesis that a species' dispersal system and three social structure produce predictable population substructure behavior. Because the same observers were monitoring within groups (5, 6). For example, adult males within groups groups on successive days, however, mating data were available of baboons and many other cercopithecine primates are pre- for only a subset of estrus cycles and usually for only one or two dicted to be less closely related than are adult females, and of the five most fertile days within any cycle. relatedness should be greater within than between matrilines. Models Predicting Paternity. For 1985-1988, we used be- The study was conducted on a group of individually known havior to predict paternity in two ways: (i) using mating wild savannah baboons, Papio cynocephalus, in Amboseli, behavior and (ii) using the priority-of-access model. For this Kenya (7). Like most cercopithecine primates and many other period, we not only had behavioral and demographic data but mammals (5, 8-11), these baboons live in polygamous multi- we were able to obtain blood samples for 27 offspring as well male, multifemale social groups in which females are matri- as for older baboons, including all adult males. To test vari- local and males disperse from their group of birth as they reach ability in reproductive success predicted by relative mating adulthood. Among baboons, close relatives do not disperse behavior, the predicted distribution was determined by using preferentially to the same group (12, 13). Within each social the subset of cycles for which conception occurred and for group, both males and females can readily be ordered in which we had observed mating behavior on at least one of the aggression-submission hierarchies in which all adult males rank fertile days (based on a 5-day fertile period). To test the above the much smaller adult females. Although the female priority-of-access model, we first determined the frequency of hierarchies are predominantly stable within and even between fertile-period overlap among females for the 120 cycles that generations, a male baboon's rank changes frequently, gener- occurred during 1985-1988. A male can monopolize only one female at a time, and the expected paternity distribution was to the The publication costs of this article were defrayed in part by page charge therefore calculated by assigning paternity solely top- payment. This article must therefore be hereby marked "advertisement" in ranking male if a conceiving female had no overlap between accordance with 18 U.S.C. §1734 solely to indicate this fact. her fertile days and those of any other female, assigning it 5797 Downloaded by guest on September 27, 2021 5798 Population Biology: Altmann et al. Proc. Natl. Acad. Sci. USA 93 (1996) equally to the top two males when two females overlapped, and Table 1. Characteristics of the human microsatellite and protein so on (2). loci used in this study Genotyping. Blood samples were collected for the genetic Locus (human analyses from 76 animals that were immobilized by an anes- map position) Allele Frequency thetic-bearing dart propelled from a blowpipe (18). The sam- ples included 54 adult females and young from Lodge Group, D2S141 128 0.81 all adult males from Lodge Group, and 17 additional adult 130 0.19 males from other nearby groups; 5 of the 6 males that had been D4S431 212 0.03 in the group in 1984 when studies began were still there in 1989 214 0.74 and could be sampled. Samples were immediately cooled and 216 0.23 within a few hours were spun and frozen for shipping. DNA was extracted from blood as described (19). Genomic DNA D6S271 166 0.08 was diluted to 1:10 for PCR reactions. Markers included 10 168 0.39 polymorphic microsatellite primer pairs at human map posi- 172 0.36 tions D2S141, D4S431, D6S271, D6S311, D7S503, DllS925, 182 0.08 D13S159, D16S402, D16S420, and D17S791 (obtained as 186 0.02 Human MapPairs from Research Genetics, Huntsville, AL) 190 0.04 (20,21) and transferrin and albumin protein loci (Table 1). The 198 0.03 PCR protocol was as follows: the forward (5') primer of each pair was end-labeled with [y-32P]ATP usingT4-polynucleotide D6S311 228 0.65 kinase (New England Biolabs) and manufacturer's buffer (70 230 0.11 mM Tris-HCl, pH 7.6). All PCR reactions were carried out in 232 0.11 a total of 10 /xl containing the following: genomic DNA diluted 234 0.13 1 in 10; 140 ,xM dNTPs; 10% DMSO; 1 mM MgCl and 0.45 D7S503 156 0.42 units Taq DNA polymerase with NH4 buffer [160 mM 158 0.06 (NH4)2SO4, 670 mM Tris-HCl, pH 8.8 at 25°C, 0.1% Tween-20] 160 0.02 with the volume made up to 10 ,ul with distilled water. PCR 164 0.02 amplifications were: one 3-min denaturation at 95°C; 35 cycles 166 0.05 of 45 s at 95°C, 1 min at two annealing temperatures (7 cycles 170 0.35 at 50°C and 28 cycles at 54°C), and 90 s at 72°C. A 10-min final 172 0.08 extension step at 72°C was also included. PCR reactions were carried out in a Hybaid Omnigene thermal cycler. Product (4 D11S925 194 0.45 pl) was loaded onto 6% denaturing acrylamide gels. The size 196 0.35 marker was M13 polycloning site sequence, using adenine and 198 0.20 thymine fragments only, loaded in one lane to create a molecular ladder.